592 William Patten 



38 cm.). In tliis manner a circular membrane — membrana circu- 

 laris — is formed, composed of nearly concentric, circular fibres, often 

 expanded into little knob-like swellings, which stain in haematoxylin, 

 and have the appearance of nuclei. It is possible that they may be the 

 aborted nuclei ofthose cells, by the fibrous trausformation of which the 

 circular membrane was formed ; but the structures in question were so 

 minute that it was impossible to arrive at any definite conclusion. The 

 circular membrane seems to be divided into two, nearly equal zones, of 

 which the inner is thicker and better defined, and terminates sharply at 

 the edge of the disc-shaped layer of rods ; the outer zone diminishes in 

 compie te ness towards the periphery, where it gradually disappears. 



The four layers just described, the outer and inner gangiionic 

 layers, the retinophorae, and the rods, are all modifications of a single 

 layer of cells forming the outer wall of the optic vesicle. Its inner wall 

 is likewise composed of four layers ; the vitreous network, the inner 

 and outer argentea, and the tapetum. In each wall of the optic ves- 

 icle, three of the four layers are cellular, while one, formed by the rods 

 in one instance and the vitreous network in the other, is non-cellular. 



The vitreous network is extraordinarily difficult to preserve in 

 its natural shape, and has, for that reason, been overlooked by previous 

 authors. Carrière (p. 104) speaks of the rods as being immersed in a 

 fatty substance (staining deeply in osmic acid) , which extends between 

 the rods as far as the sieve-membrane on the one side, and is limited by 

 the tapetum (argentea) on the other. This fatty mass is formed by the 

 fusion of the sheaths of the poorly preserved rods, while what he con- 

 sidered as the rods are simply the axial cores. The fatty mass of 

 Carrière must not, therefore, be confouuded with the vitreous net- 

 work, which may be regarded as a very thin layer of hyaline sub- 

 stance, perforated by large holes into which the inner ends of the rods 

 fit (PI. 29, fig. 21). The holes, then, are as large, and have the same 

 hexagonal or pentagonal shape, as the inner ends of the rods ; they are 

 separated from each other by narrow bars, provided at their points of 

 union with short, vertical projections, which ascend a short distance be- 

 tween the inner ends of the rods. — The vitreous substance, therefore, 

 forms a complete network, the meshes of which constitute a crown for 

 the inner end of each rod, completely filling the narrow space left be- 

 tween their rounded edges and the argentea. On the periphery of the 

 retina, where the rods are absent, the vitreous network is transformed 

 into a thin piate filled with numerous and irregulär holes (PI. 29, 

 fig. 21 «). The vitreous network, as is indicated by its development, is 



